advance sed to remove multile lines.

ufmale · 05-12-2010, 03:25 PM

i have a txt file with couple of comment lines:

Number of title = !num!
#line1
#line2
#line3

I wrote a script with "sed" to replace !num! in this file, which
is very straightforward. However, based on the !num!,
i want to remove the number of "#" based on the !num! value.

is there an easy way to do that with "sed"; otherwise, i will have to write a script to loop through the file.

MTK358 · 05-12-2010, 04:22 PM

Quote:

Originally Posted by ufmale

i have a txt file with couple of comment lines:

Number of title = !num!
#line1
#line2
#line3

I wrote a script with "sed" to replace !num! in this file, which
is very straightforward. However, based on the !num!,
i want to remove the number of "#" based on the !num! value.

is there an easy way to do that with "sed"; otherwise, i will have to write a script to loop through the file.

I think this will be much easier and more practical in AWK.

I don't understand this !num! thing, though — is the literal string "!num!" actually in the file, or is it a number? And what do you mean by "replacing" it?

pixellany · 05-12-2010, 04:24 PM

Post a sample of the data before and after the desired changes.

ufmale · 05-12-2010, 10:18 PM

sorry for unclear.. The desired result is that if i provide 2 to replace !num!,
then the result would be

Quote:

Number of title = 2
line1
line2
#line3

If i provide 3, then

Quote:

Number of title = 3
line1
line2
line3

grail · 05-13-2010, 12:21 AM

Code:

awk '/^Num/{cnt=$NF}cnt{sub(/^#/,"");print;cnt--}' file

MTK358 · 05-13-2010, 07:48 AM

Code:

BEGIN {
	rmlines = 0
}

{
	if ($NF == "!num!")
	{
		$NF = num
		rmlines = num
		print
	}
	else if ($0 ~ /\#.*/)
	{
		if (rmlines == 0)
		{
			print
		}
		else
		{
			rmlines--
		}
	}
	else
	{
		print
	}
}

grail · 05-13-2010, 08:08 AM

@MTK358 - I cannot see where you remove the #?

MTK358 · 05-13-2010, 08:11 AM

Oh, I thought I should remove the whole line.

<comment deleted>

MTK358 · 05-14-2010, 09:44 AM

I think this should do it (untested):

Code:

BEGIN {
	rmlines = 0
}

{
	if ($NF == "!num!")
	{
		$NF = num
		rmlines = num
		print
	}
	else if ($0 ~ /\#.*/)
	{
		if (rmlines == 0)
		{
			print
		}
		else
		{
			print substr($0, 2, length($0))
			rmlines--
		}
	}
	else
	{
		print
	}
}